n8n is a workflow automation tool. You connect a trigger to a series of steps, and it runs that sequence whenever the trigger fires. If you have used Zapier or Make, the concept is familiar. n8n differs in two ways. You can run it on your own server, and its steps are closer to code than to a form.
Quick answer: n8n is an open-source workflow automation tool that connects your apps, databases and AI models with a visual editor and code when you need it. You can self-host it for free or use n8n Cloud. It charges per workflow run rather than per step, so multi-step automations cost less than they do on Zapier.
This guide covers what n8n does, where it differs from the alternatives, and the questions worth answering before you commit to it.
How an n8n workflow is put together
Every workflow starts with a trigger, such as a webhook, a schedule, a new row in a database or an incoming email. After that you add nodes. Each node does one thing: call an API, transform data, branch on a condition, wait, or run a snippet of JavaScript.
The canvas is visual, but the data moving between nodes is plain JSON. When a workflow fails, you can open any node, see what came in and what went out, and re-run from that point. You debug by reading the data at each step.
- Triggers start the run: webhooks, cron schedules, app events or manual execution.
- Action nodes send HTTP requests and talk to databases, CRMs, spreadsheets and messaging platforms.
- Logic nodes branch, loop, merge and filter. Most of a workflow's complexity sits in these.
- Code nodes let you drop into JavaScript or Python when no node does quite what you need.
Code nodes cover what the other nodes cannot. Most automation platforms eventually hit a transformation they cannot express. In n8n you write four lines of JavaScript and move on.
Self-hosted or cloud
n8n offers a hosted cloud product and a self-hosted option. The self-hosted version runs in Docker and is free to use under a fair-code licence, which permits internal business use but restricts reselling it as a competing hosted service.
Self-hosting is worth considering when your workflows touch sensitive data, when you don't want a per-execution cost, or when n8n has to sit inside a private network to reach internal systems. It still costs time. Somebody has to patch the server, watch the queue, manage backups and handle upgrades, and that work never stops.
Cloud makes sense when you want n8n to run the server for you and your volumes are predictable.
How it compares to Zapier and Make
Zapier is the easiest to start with and the most expensive to scale, because it prices per task and most of its power sits behind higher tiers. Make sits in the middle. It has a visual canvas with more logic than Zapier and a lower cost per operation.
n8n is slower to get started with and gives you more control. You get self-hosting, code nodes, and no per-task pricing if you run it yourself. You also take on more maintenance. If your automations are simple and your volume is low, Zapier will cost less in total once you count your own time. If your automations branch, touch internal systems, or run at volume, n8n is usually the better fit.
Exceptions and failure handling
A workflow that handles the happy path is perhaps a third of the work. The rest is what happens when the API times out, the record is missing a field, someone enters a date in the wrong format, or the same webhook fires twice.
Before building anything in n8n, write down what should happen when each step fails. Who gets told? Does the workflow retry, and how many times? Is there a manual override for the person who has to fix it at 5pm on a Friday? n8n supports retries, error branches and dedicated error workflows, but you have to design them in.
Plan the failure paths before you open the canvas.
When n8n is the wrong answer
If you need one simple connection between two mainstream apps, a purpose-built integration or Zapier will be quicker to set up.
If nobody on your side can maintain a server or read a stack trace, self-hosting will become a liability. Use the cloud version or pick a managed tool.
If the people who perform a process do not agree on how it works, automation will encode the disagreement and run it faster.
Questions worth answering first
- What triggers this process today, and who currently decides it should start?
- Which systems hold the data, and who owns the credentials?
- What are the known exceptions, and how are they handled now?
- Who needs to be told when something fails, and through which channel?
- Who maintains this in six months when the person who built it is busy?
If you can answer those five, the build is straightforward. If you cannot, the build may still be quick, but the workflow will break easily.
Frequently asked questions
It connects business tools so repeated work runs automatically. Typical uses are lead follow-up, quoting, invoicing, reporting, support triage, data sync and AI agents that read, classify and reply.
Written by
Founder & CEO of Wasevo. Builds SEO, software and AI automation for clients in the US, UK, Canada, Australia, Sweden and Pakistan since 2020.
